Build an end-to-end fiber latency budget from route distance, propagation speed, active-device delay, processing time, queueing, and application overhead. Compare the estimated one-way delay and round-trip time with a user-defined RTT target to reveal remaining margin.
A positive target margin means the modeled RTT is within the target. A negative margin identifies the latency reduction required, but it does not reveal which component can realistically be changed.
Actual paths can be longer than mapped distance and queueing can vary sharply. Measure production paths and include protocol handshakes separately when they matter to user experience.
An 800 km route at 200,000 km/s adds 4 ms one way. Eight devices at 0.08 ms, plus 1.5 ms processing and 2 ms queueing, create 8.14 ms one-way network latency. With 4 ms application overhead, estimated RTT is 20.28 ms.
Light travels more slowly through glass than through vacuum, so fiber designs commonly use an approximate propagation speed near 200,000 km/s.
Enter one-way cable-route distance. The formula doubles the one-way network latency when calculating RTT.
Yes. Device delay is included in the one-way budget, and the complete one-way network delay is doubled for round trip.
It is delay already expressed at the round-trip or transaction level, such as extra processing not represented in the one-way network components.
Only if the queueing allowance represents that condition. Real congestion is variable, so measured distributions are preferable for operational planning.
| Input | Purpose |
|---|---|
| One-way fiber route length | User-adjustable planning assumption |
| Propagation speed in fiber | User-adjustable planning assumption |
| Active network devices | User-adjustable planning assumption |
| Delay per active device | User-adjustable planning assumption |
| Endpoint processing, one way | User-adjustable planning assumption |